Mar Biol DOI 10.1007/s00227-007-0769-y RESEARCH ARTICLE First age and growth estimates in the deep water shark, Etmopterus Spinax (Linnaeus, 1758), by deep coned vertebral analysis Enrico Gennari · Umberto Scacco Received: 2 May 2007 / Accepted: 4 July 2007 © Springer-Verlag 2007 Abstract The velvet belly Etmopterus spinax (Linnaeus, by an alternation of translucent and opaque areas (Ride- 1758) is a deep water bottom-dwelling species very com- wood 1921; Urist 1961; Cailliet et al. 1983). Vertebral mon in the western Mediterranean sea. This species is a dimensions, as well as their degree of calciWcation, vary portion of the by-catch of the red shrimps and Norway lob- considerably within the elasmobranch group (La Marca sters otter trawl Wsheries on the meso and ipo-bathyal 1966; Applegate 1967; Moss 1977). For example, vertebrae grounds. A new, simple, rapid, and inexpensive vertebral of coastal and pelagic species are more calciWed than those preparation method was used on a total of 241 specimens, of bottom dwelling deep-water sharks (Cailliet et al. 1986; sampled throughout 2000. Post-cranial portions of vertebral Cailliet 1990). These diVerences are also reXected in varia- column were removed and vertebrae were prepared for age- tions of shape and in growth zone appearance, such as the ing readings. Band pair counts ranged from 0 to 9 in presence and quality of bands and/or rings. Due to these females, and from 0 to 7 in males. Von BertalanVy growth diVerences, a general protocol for the elasmobranch group equations estimated for both sexes suggested a higher is not really available because of the high variability of cal- W longevity for females (males: L1 = 394.3 mm k =0.19 ci cation degree among species (Applegate 1967; Cailliet W t0 = ¡1.41 L0 = 92.7 mm A99 = 18.24 years; females: L1 = et al. 1983). Based on identi cation and count of band 450 mm k =0.16 t0 = ¡1.09 L0 = 72 mm A99 = 21.66 years). pairs, many techniques have been developed to assess age Age estimations are discussed in the context of deep water in cartilaginous species, assuming an annual periodicity shark species. This is the Wrst successful attempt at delin- (Cailliet et al. 1990). Even if deposition time should be eating faint growth bands in the poorly calciWed deep coned properly validated within each species (Beamish and vertebrae of E. spinax. This technique may be used in other McFarlane 1983; Cailliet 1990; Campana 2001), age esti- diYcult poorly calciWed species. mates have not always been validated (Campana 2001). Etmopterus spinax is a bottom dwelling shark, typical of the bathyal stratum (200–1,000 m depth) of the Mediterra- Introduction nean sea (Tortonese 1956; Fischer et al. 1987; Notarbartolo Di Sciara and Bianchi 1998). It is caught as by-catch by In cartilaginous Wsh, hard and calciWed structures, such as commercial otter trawlers (Bertrand et al. 2000; Relini et al. spines and vertebral centra, exhibit growth zones, identiWable 2000). E. spinax is characterized, as well as other deep bot- tom dwelling sharks (Etmopterus baxteri, Galeus melasto- mus, Galeorhinus galeus), by typical amphicoelous “deep coned” vertebrae; additionally its vertebrae are very small (Fig. 1), have little calciWcation and outer margins tend to refold on themselves as Wsh size increases (Fig. 11). There- Communicated by R. Cattaneo-Vietti. fore, these features, together with the faint markings and fragile condition of the vertebrae, render them unsuitable to E. Gennari · U. Scacco (&) I.C.R.A.M, Via di Casalotti 300, 00166 Rome, Italy both most traditional ageing and validating increment peri- e-mail: [email protected] odicity techniques. The simple, inexpensive, and rapid 123 Mar Biol Fig. 2 Trawl grounds oV the Latium coast where Etmopterus spinax specimens were sampled Fig. 1 Typical vertebral size compared to an observer’s Wnger. Scale bar 1 mm water, it was dipped again in a new but same concen- trated NaClO–water solution, for the same amount of method proposed in this work, modiWed from Hoenig and time, and so on until complete removal of connective Brown (1988), produced band elucidation for E. spinax ver- and muscular tissues, and separation of each single ver- tebrae, leading to the Wrst, but unvalidated, age results for tebra. this species. 3. To measure the vertebral length (VL), each vertebra was put under transmitted light on a stereo-microscope provided with digital still camera, and millimeter paper Materials and methods for calibration. A digital image was acquired, and mea- surements were taken using Image-Pro©Plus. Due to Samples were obtained from commercial otter trawl opera- the small sizes of these vertebrae, VL, rather than tions, conducted by Fiumicino marine, oV the Latium coast vertebral diameter, was chosen to analyze vertebral (central Tyrrhenian sea) (Fig. 2). A combined eVort of 16 growth. hauls distributed in 3 bathymetrical strata (300–490 m, 4. Each centrum was dipped into a 5% in volume W 491–580 m, and 581–900 m) was made seasonally during Co(NO3)2–H2O solution in order to stain calci ed 2000. A total of 241 specimens of Etmopterus spinax areas. Time of immersion varied between 1 and 5 min, (Linnaeus 1758) were collected and dissected. Each speci- depending on vertebral dimension and hence on its men was sexed (152 females and 89 males), measured (over degree of calciWcation. Each test tube had to be stirred the body total length and pre-caudal length, TL and PL gently in order to allow Co(NO3)2penetration into the respectively) to the nearest millimetre, and weighed (evis- pronounced cavities of the deep-coned vertebra. cerated weight, WE) to the nearest centigram. According to 5. Vertebral centrum was rinsed with distilled water for the estimates of velvet belly size at Wrst maturity in the few seconds in order to remove excess Co(NO3)2. Mediteranean sea (Vacchi and Relini Orsi 1979; Fischer 6. To enhance the bands, each centrum was dipped into an et al. 1987), specimens were divided into 3 size classes: alcohol acid–water solution (obtained adding hydro- “juveniles” (I: 100–200 mm TL, n = 56), “subadult” (II: chloric acid into a 70% in volume ethanol–water solu- 201–301 mm TL, n = 139), and “adult” (III: 301–435 mm tion, with a ratio of 1:20) between few seconds to one TL, n = 46). Sex ratio was calculated for size class as fol- minute, depending on vertebral dimension. Time of low: Females/(Females + Males) 100. A small post-cranial immersion was the most critical step of the entire prep- section of vertebral column (3–4 vertebrae) was removed aration: a shorter time could not allow band enhance- from each specimen. ment, whereas a prolonged dipping could destroy the Vertebrae were prepared according to the method pro- centrum. posed by Hoenig and Brown (1988), but modiWed for this 7. After being rinsed using distilled water, the stained study as follow: centrum was then viewed under the transmitted light using a stereo-microscope, and a digital image was 1. After removal, vertebral column section was stored at obtained and read independently by two readers. ¡20°C. 2. Section was dipped in a 10% in volume NaClO–water The criterion chosen for band readings is based on the iden- solution for 15 min. After being rinsed using distilled tiWcation of a dark layer followed by a lighter one deWned 123 Mar Biol as a band pair (Cailliet et al. 2006). Each band was consid- DiVerence in percent of occurrence between type errors ered a temporal growth zone (Fig. 3). (§1 and §2) was tested by a Chi-square test. Each reader counted, twice for each centrum, dark bands Estimates of precision were evaluated following the on the vertebral outer surface, from the core toward the dis- methodology suggested by Goldman (2004): total percent- tal margin (EA = estimated age). Counting was performed age agreement (PA), percentage agreement plus-minus one by each reader without knowledge either of the specimen band (PA § 1) and percentage agreement by size classes length or the other reader’s count. PA (I, II, III) were calculated both between and within readers. Chi-square test on contingency tables was utilized Statistical analysis to check for bias in all these cases. The age-bias curve (Campana et al. 1995) was utilized to test for bias between DiVerences in sex ratio among sizes classes were tested by readers within age groups. In addition the non-parametric a chi-square Test on a 2 £ 3 contingency table. Wilcoxon Test (Conover 1971), index of average percent Linear regression analysis was used to calculate length– error (IAPE), and mean coeYcient of variation (CV) weight relationships (total length on transformed natural (Beamish and Fournier 1981) were calculated in order to logarithm of eviscerated weight), as well as the VL-TL, and provide further estimates of precision in band count estima- EA-TL correlations. A natural logarithmic transformation tion between readers. The statistical package Statistica (6.0) of (EA + 1) was also needed due to the presence of “0” val- 2006 was used for the aforementioned analyses. ues. ANCOVA analysis was then used to investigate sex The von BertalanVy growth parameters were determined diVerences for all the aforementioned relationships. for each sex using the Wshery program FISHPARM (Prager ¡k(t ¡ t0) et al. 1987), using the equation L(t)=L1(1 ¡ e ) (BertalanVy 1960). Longevity was estimated through the algorithm A99 = 5·Ln(2)/k (Fabens 1965), where A99 is the time (in years) passed before reaching 99% of L1 and k is the growth constant derived from the von BertalanVy growth equation. Extrapolated longevity, based upon A99, is believed to give a more realistic estimate of the maximum age rather than A97 and A95 (Skomal and Natanson 2003).